Irene Pagliai

RTG PhD student, member since 2021

Project "Identifying context-dependent meaning components of figurative expressions"

My thesis project focuses on figurative language, more precisely on Italian and English idioms. Idioms are known to be complex linguistic units whose meaning is not compositional; for example, both "break the ice" in English and "rompere il ghiaccio" in Italian have the figurative meaning of “overcoming an initial moment of awkwardness in interpersonal relationships”. However, in some exceptional cases, the figurative meaning may interact with the literal one, which remains exceptionally relevant in context. See for instance (1):

(1) [ITA] Quanti anni…" gli dissi tanto per rompere il ghiaccio. Proprio di ghiaccio si trattava… Era più freddo di un iceberg!

[ENG] ‘How many years..." I told him just to break the ice. Just ice it was - he was colder than an iceberg!’ [CORIS]

Some idioms with specific characteristics are supposed to be particularly prone to such ambiguous uses. For this reason, the project is divided into two parts: the first one is devoted to the study of idioms as types (i.e., without context); to collect data regarding the relevant features, a lexicon of cross-linguistic idioms will be created exploiting crowdsourcing. The second part will focus on the analysis of idioms as tokens (i.e., in ambiguous contexts such as the one in the example). Both parts will be necessary to try to understand a) whether it is possible to identify cross-linguistic patterns of contextual ambiguity that best fit the characteristics of idioms b) whether idioms with a specific combination of characteristics are by and large more compatible than others with such ambiguous contexts.

Supervisors: Anke Holler , Alessandro Lenci , Caroline Sporleder


I received my Bachelor's degree in Humanities from the University of Bologna, with a thesis in sociolinguistics regarding the importance of considering the many social factors impacting language when teaching Italian as L1. I continued my studies in Bologna, where I completed my Master's degree in Linguistics; during the same time, I spent one year at Jean Jaurès University in Toulouse, France. My Master's thesis focused on the construction of the Italian FrameNet, a computational lexicon exploitable by Natural Language Processing techniques.


My approach to linguistics is usage-based and functional in nature. At the level of theoretical analysis, my focus is on lexical semantics, on multiword expressions and the idiom principle. As I consider language an integral part of our cognitive faculties, I believe it is crucial to include in the picture psycholinguistic analyses as well. Broadly speaking, my approach to the discipline has always been a multidisciplinary one: Natural Language Processing can also by included in my interests, as I think it is important to grasp how the latest generation of linguistic models are revolutionizing the field and the role of the linguist. Staying up to date on this front is also important to be able to face new challenges, such as the presence of important socio-cultural biases within neural models. Moreover, this theme also connects with the key role that I believe sociolinguistics has in describing the complex circuit of mutual influence between language and society.